Human GWAS data users

The Independent Data Access Committee has approved access to MalariaGEN human GWAS data for the following researchers. Click on a link below to learn more about the researchers and how they are using this data:

 


 

Data Users: Prof. Alkes Price (PI), Dr Samuela Pollack, Dr Nick Patterson, D. David Reich, Prof Pardis Sabeti and Dr Noah Zaitlen

Institution: Harvard University

Project: Identifying malaria resistance genes via natural selection mapping

The study of recent natural selection in human populations has important applications to human history and medicine. Positive natural selection drives the increase in beneficial alleles and plays a role in explaining diversity across human populations. By discovering traits subject to positive selection, we can better understand the population-level response to environmental pressures including infectious disease. Our study examines unusual population differentiation between three large data sets to detect natural selection. The populations examined, African Americans, Nigerians, and Gambians, are genetically close to one another, and because there are not many genetic differences between them, when there is a difference found it is an individuation of something potentially very significant. The results of this study imply that among these three closely related populations, natural selection is at work in genes related to malaria, and bladder and gastric cancer. Understanding how genes affect disease susceptibility can help identify new treatments and improve existing treatments. In this study the results are significant because they can help broaden understanding of how populations respond genetically to environmental pressures such as infections disease.

 


 

Data Users: Prof. Andreas Ziegler

Institution: University of Luebeck

Project: Powerful and robust association tests and model selection for unrelated individuals and family-based samples in the framework of the generalised linear model

The analysis of genome-wide genetic association (GWA) studies generally starts with univariate statistical tests of each single nucleotide polymorphism. The standard approach is the Cochran-Armitage trend test which can lose considerable power if the true genetic model is not additive.

An alternative is the MAX test which is robust against the three basic modes of inheritance. We have derived the asymptotic distribution of the MAX test using the generalised linear model together with the Delta method and multiple contrasts. It allows us to test and estimate genetic effects and the selection of the most plausible genetic model. Recently, we applied the model selection approach to a GWA study on severe falciparum malaria (Timmann et al. 2012, Nature).

In brief, 1325 severe malaria cases and 828 unaffected controls from Ghana were genotyped on the Affymetrix Genome-Wide Human SNP Array 6.0. Logistic regression was fitted with adjustment for age, gender, and population stratification, the latter using a principal component analysis.

We identified 102 SNPs located in 41 distinct genomic regions using thresholds of p<5xE-5. The most likely mode of inheritance was identified using the MAX test model selection approach. Replication was performed subsequently in an additional 1320 severe malaria cases and 2222 controls from the same population using the most likely mode of inheritance.

An independent replication group was made available through data provided by the MalariaGEN Network of a GWAs on 958 severe malaria cases among Gambian children and 1382 controls, a different ethnic group, which was genotyped on the AffymetrixGeneChip Human Mapping 500K Array Set.

Finally, two novel loci were genome-wide significant (p<5xE-8), and the first has lead SNP rs10900589 on chromosome 1q32.1 within the ATPase, Ca++ transporting, plasma membrane 4 (ATP2B4) gene. The most likely mode of inheritance for the initial Ghanaian GWA sample and the combined Ghanaian samples is a recessive one with an effect of the risk allele of 0.58 with a minor allele frequency of 0.36. The positive predictive value for this choice is almost 100%.

Replication in the data from MalariaGEN was successful only for the recessive model.

Publication: Loley Ch., Koenig I. R., Hothorn L, Ziegler A. A unifying framework for robust association testing, estimation, and genetic model selection using the generalised linear model. European Journal of Human Genetics 2013:  1-7.

 


 

Data Users: Prof. Anne E Hughes (PI) and Dr Declan T Bradley

Institution: Queen's University of Belfast

Project: Genetic susceptibility to common immune-related disorders

We will investigate whether variation of people's genome in a basic and central part of the immune system called the complement system affects their susceptibility to malaria. Variations in these genes affect susceptibility to auto-immune and inflammatory diseases such as age-related  macular degeneration, as well as to other infections such as meningitis and septicaemia due to a bacteria called Neisseria meningitidis. We will use publicly available population genomic data from African populations to understand the nature and frequency of variations in these genes in African populations and use this knowledge to find out whether their frequency differs between people with and without Malaria in the MalariaGEN genetic study.

 


 

Data Users: Dr David Torrents Arenales (PI) and Dr Josep Maria Mercader

Institution: Barcelona Supercomputing Centre

Project: Systems biology approaches to assess the combined effect of multiple SNPs on susceptibility to malaria by multiple logistic regression models

Genome-Wide Association Studies (GWAS), have allowed the identification of new unsuspected variants associated to complex diseases, and also to the susceptibility of infectious diseases. However, how these variants increase the risk of a certain diseases is still unclear. The aim of this project is to apply different GWAS methodologies to identify novel key genes for the susceptibility of malaria infection. These new approaches include systems biology techniques, novel gene-based tests, and new pathway analyses methods. We also plan to integrate the results with other available genomics data related to malaria infection in order to provide a better interpretation of the GWAS findings.

 


 

Data Users: Dr Desmond Smith (PI) and Dr Andy Lin

Institution: University of California  

Project: Two-way and three-way genetic interaction that underlie human diseases

We are exploring computational strategies to explore the effects and consequences of gene-gene interactions in humans. Unfortunately, incorrect associations can arise in genetic analyses due to hidden population structure. However, association tests based on parents and their children (trios) can avoid the complications of population structure and so are ideal for mixed populations. We will use the MalariaGEN data to develop trio-based tests for genetic interactions, identify two-locus interactions associated with malaria using exhaustive searches, evaluate the performance of several strategies to enhance computational efficiency in uncovering genetic interactions, and develop tests for identifying three-way genetic interactions and estimating the sample sizes needed to detect three-way interactions.

 


 

Data User: Prof. Heather Cordell

Institution: University of Newcastle

Project: Investigating whether multiple regression strategies can outperform imputation for detecting association with causal variants hat are poorly tagged by SNPs on genotyping arrays

 


 

Data Users: Dr James F. Wilson (PI) and Peter Joshi

Institution: University of Edinburgh

Project: Individual genome-wide homozygosity-distribution and relationship to disease

We investigated the association between susceptibility to malaria and the degree to which the genome is made up of stretches of identical DNA inherited from both parents (runs of homozygosity). Homozygosity has been posited to influence infectious disease risk but we saw no significant relationship.

 


 

Data User: John Farrell

Institution: Boston University  

Project: Data mining genome-wide association study results for HLA disease

The MalariaGEN genome scan data is being analysed to identify HLA genotypes that may influence susceptibility to malaria. A novel method to impute HLA alleles from genome scan SNP data in the MHC region has been developed. This new method will now be used to test association of severe malaria with these imputed HLA alleles.

 


 

Data Users: Prof. Matthew Stephens (PI) and Dr Bryan Howie

Institution: University of Chicago    

Project: Statistical issues in association studies of African populations

“Genotype imputation” is a statistical technique that can help identify genetic risk factors for disease. The idea is to use near-complete genetic information from a small number of people to predict the DNA sequences of many other people recruited for a genetic association study. By providing a more complete picture of the genetic variation carried by study participants, genotype imputation methods improve our ability to find parts of the genome that my affect risk for disease.

For genotype imputation to work, we need to have a “reference panel” of detailed genetic information to drive the predictions in our study. A typical reference panel might include the genome sequences of hundreds of people. Reference panels have now been generated for human populations around the world, and a key question in any disease study is which reference panel will most accurately predict the genomes of study participants.

This can be a hard question to answer because of the complex historical relationships between human populations. To address this problem, we developed a method that can automatically choose the best reference panel for each person in a study without needing to know about population history. Our approach makes it easy for disease investigators to accurately predict the genomes of study participants, and it is especially useful in African populations because their complex histories make it hard to identify the best reference panels by other means. This work will accelerate the discovery of genetic risk factors for disease and ultimately lead to better treatments.

Publications: Howie B, Marchini J, Stephens M. Genotype imputation with thousands of genomes. G3: Genes, Genomes, Genetics 2011, 1(6): 457-470.

 


 

Data Users: Prof. Michaël Guedj (PI), Dr Daniel Cohen, Dr Ilya Chumakov, Dr Serguei Nabirotchkine, Ms Claudia Satizabal, Mrs Fabrice Glibert, and Mrs Jonas Mandel

Institution: Pharnext

Project: Replication in genome-wide association studies: from replication of SNPs to replication of gene-networks

Pharnext SAS is a French privately-held biopharmaceutical company of 30 people, founded in April 2007 by Prof. Daniel Cohen and his main collaborators. Pharnext is the pioneer of a new paradigm in R&D, which is bound to revolutionise the way diseases are treated based on a fundamental concept, the Pleotherapy. Pleotherapy aims at identifying the best combination of active molecules in order to restore the molecular pathways perturbed in each disease and addresses the shortcoming of the standard R&D approach.

The traditional “one drug, one disease” paradigm, under which a single drug is used to treat a single yet often multi-factorial disease, has indeed shown its limits in terms of efficacy and safety. The novel approach led by Pharnext allows targeting several molecular ‘nodes’ in disease-perturbed pathways and thus helps to increase the treatment’s efficacy and safety. It consists in combining mini-doses of several drugs already approved by healthcare authorities for other diseases into a single patented compound, a pleodrug.

Pharnext intend to develop validation steps based on genetics to strengthen the validity, homogeneity and robustness of the identified targets.

The aim of the project is to use the genome-wide genetic profiles to consider malaria in the context of the Pharnext research process.

 


 

Data Users: Prof. Michael Levin (PI), Dr. Victoria Wright, Dr. Lachlan Coin, Dr. Hariklia Eleftherohorinou, and Dr. Clive Hoggart

Institution: Imperial College London

Project: Biological pathway analysis of malaria genome-wide association study

We aim to undertake a pathway-based analysis of the MalariaGEN data using the methodology for pathway analysis developed by the Imperial College team (Eleftherohorinou et al, 2009). Pathway analysis of GWAS provides new insights into genetic susceptibility to 3 inflammatory diseases (PLoS One, 4(11), e8068). Rather than analysing genotypic data at a single SNP level, pathway-based analysis has proved to be a powerful new method for identifying key biological pathways involved in disease. The method may be less affected by confounding effects of ethnic differences than single SNP analysis, and has proved a powerful new approach to identifying biologically important pathways in other genetic data sets.

 


 

Data Users: Prof. Patrick Chinnery (PI) and Dr. Gavin Hudson

Institution: University of Newcastle

Project: Variation in the human transcriptional program between uncomplicated and severe malaria infection provides insight into potential gene candidates for investigation into malaria susceptibility

There is emerging evidence that genetic variation in maternally inherited mitochondrial DNA influences survival, particularly in the context of infection.We are studying the role of common mtDNA variants genotyped as part of the MalariaGEN study.

 


 

Data Users: Prof. Rolf Horstmann (PI), Dr K Schuldt, Dr med. C Timmann, Dr med. T Thye

Institution: University of Hamburg

Project: Human genetic resistance to mild and severe malaria

Our project aims to identify human genetic resistance marker for mild and severe malaria. By comparing large groups of patients and controls, which are differently affected, genetic variants are being identified. These variants are studied with regard to their implication for disease and disease severity. The functional effects of disease-associated genetic variants are implied to be involved in defence mechanisms or other metabolic pathways relevant to the course of infection. Thereby, they can reveal previously unrecognised approaches to treatment and prevention.

Publications: Timmann et al. Genome-wide association study indicates two novel resistance loci for severe malaria. Nature. 2012 Sep 20;489(7416):443-6. doi: 10.1038/nature11334. Epub 2012 Aug 15. PMID: 22895189.

 


 

Data Users: Prof. Steven McCarroll (PI) and Ms Linda Boettger

Institution: Harvard University

Project: Human genome copy member variation and malaria susceptibility

 


 

Data Users: Dr Steven Taylor (PI) and Dr Barbara Engelhardt

Institution: Duke University

Project: Severe malaria and mutations in tmprss6

We are taking a candidate-gene approach to identifying mutations associated with protection from severe falciparum malaria. We are interested in mutations which influence erythrocyte phenotypes. Such mutations have been described by recent genome-wide studies of erythrocyte functional variants. We hypothesize that these innate red blood cell variants confer protection from severe malaria, similar in kind if not degree to that described for hemoglobinopathies or enzymopathies. Describing innate mechanisms of protection from disease can reveal mechanisms of pathogenesis that can serve as targets for future therapeutic, preventive, and adjunctive interventions.